Quantification strategies in real-time RT-PCRwith special focus on RNA quality,

relative expression and data normalisation

qPCR Satellite SymposiumBio City Leipzig, 10 - 11th March 2005

PD Dr. Michael W. PfafflPhysiology - WeihenstephanCenter of Life and Food SciencesTechnical University of MunichWeihenstephaner Berg 385350 Freising-Weihenstephan; Germanymichael.pfaffl@wzw.tum.dewww.gene-quantification.info

Genotype -> Phenotype -> Function

DNA => pre-mRNA => mRNA => Protein => Function↓ ↓ ↓

Transcription Splicing Translationpost-translatoric modifications

Genom Transcriptome Proteome Metabolome& Splicome

TranscriptomeRNA content of a cellT.A. Brown, GENOMES 2nd Edition

• ribosomal RNA rRNA 80-85% ( 5S, 18S und 28S )• transfer RNA tRNA 10-15%• messenger RNA mRNA 2% (1-5%) ( ∅ length 1930 bases)

• high abundant < 100 genes > 10,000 copies/cell• intermediate abundant ~ 500 - 1,000 genes 200 - 400 copies/cell• low abundant ~ 27,000 genes < 10 - 50 copies/cell

• RNA quantity & RNA quality

Quantification Strategies in real time qRT-PCRM.W. Pfaffl, BioSpektrum 2004 (Sonderausgabe PCR)

relative quantificationabsolute quantification

external calibration curve

one color detectionsystem

SYBR Green I

external calibration curve

two color detectionsystem

e.g. Probes, ROX

external calibration

curve withoutany referenece

gene

normalisationvia one

reference gene

via reference gene index

>3 HKG

external calibration curve

• RT-PCR product

• plasmid DNA

• in vitro transcribed RNA

• synthetic DNA Oligos

• synthetic RNA Oligos

without real-time PCR efficiency correction

2 (-∆∆ CP) REST, qGeneLC software

with real-time PCR efficiency correction

ROXROX

Quantification strategies in real-time RT-PCR

Absolute quantification using calibration curves

• recombinant DNA (recDNA) calibration curve (Pfaffl & Hageleit, Biotechnol.Lett. 2001)

• recombinant RNA (recRNA) calibration curve (Pfaffl & Hageleit, Biotechnol.Lett. 2001)

• calibration curve using a synthetic DNA oligo-nucleotide (Bustin, JME 2000)

• calibration curve using a synthetic RNA oligo-nucleotide (Bustin et al. 2000)

• calibration curve using a purified RT-PCR product (Einspanier et al. 1999)

• „Copy & Paste“ of previously performed calibration curves (LC software)

• Calibration curve using a purified RT-PCR product or a synthetic ss/ds oligo-nucleotidetwo-step RT-PCRadvantages: quick, highly defined DNA content for the synthetic oligodisadvantages: instable, “often problems with re-amplification”, „short“ templates

Absolute quantification using calibration curves

• Calibration curve using a recombinant DNA (recDNA), e.g. plasmid DNAtwo-step RT-PCRadvantages: very stabile, no problems with re-amplification, „mimic of mRNA“disadvantages: cloning, linearization and purification of recDNA

• Calibration curve using a recombinant RNA (recRNA)one-step RT-PCR => recRNA and native mRNA undergoing RT and PCR in paralleladvantages: mimics the natural mRNA situation best (recRNA = native mRNA) disadvantages: very instable recRNA, complicate cloning, linearization,

purification of recRNA, storage problems, reproducibility (???)=> storage of recRNA !!!

• „Copy & Paste“ of previously performed calibration curves (e.g. LightCycler Software)advantages: very easy and very high reproducibility (at least for the calibration curve)disadvantages: do not covers variations in real RT-PCR experiment: RNA quality, slope,

qPCR efficiency ??? batch to batch variations, etc. => truth ???

ER-alpha intra-assay variation CV = 18.7% (n = 3)

input ss cDNA molecules

1.65 e+91.65 e+81.65 e+71.65 e+61.65 e+51.65 e+41.65 e+3

dete

cted

mol

ecul

es

1e+3

1e+4

1e+5

1e+6

1e+7

1e+8

1e+9

1e+10 31.4% 14.4% 7.3% 7.9% 33.6% 13.7% 22.8%

input ss cDNA molecules

1.65 e+91.65 e+81.65 e+71.65 e+61.65 e+51.65 e+41.65 e+3

dete

cted

mol

ecul

es

1e+3

1e+4

1e+5

1e+6

1e+7

1e+8

1e+9

1e+10 41.2% 31.9% 46.2% 20.6% 15.1% 21.4% 23.8%

ER-alpha inter-assay variation CV = 28.6% (n = 7)

ERα intra-assay & inter-assay variationvariation on the basis of detected molecules using a recombinant plasmid DNA calibration curveintra-assay variation: within one runinter-assay variation: between different runs

Validation: absolute quantification of steroid receptors

25.7% (n = 4)29.7% (n = 4)28.6% (n = 4)24.3% (n = 7)inter-assay variation

83.5[ 82.9 ]

83.983.883.183.5

[ 87.9 ]89.0

[ 89.9 ]90.190.590.2

86.085.0

--85.385.486.0

85.484.485.085.5

--84.5

Species specific Tmelt (°C)Homo sapiens

Rattus norvegicusCallithrix jacchus (primate)

Bos taurusOvis ariesSus scrofa

5.7% (n = 4)17.6% (n = 4)18.7% (n = 4)31.2% (n = 3)intra-assay variation

93.9%81.3%81.2%90.7%PCR efficiency

760 – 7.60*109

molecules(r = 0.998)

106 - 1.06*1010

molecules(r = 0.996)

165 - 1.65*109

molecules(r = 0.995)

120 - 1.20*1010

molecules(r = 0.998)

quantification range(test linearity)

760 molecules106 molecules165 molecules120 moleculesquantification limit

14 molecules10 molecules2 molecules12 moleculesdetection limit227 bp262 bp234 bp172 bpproduct length

PRERβERαAR

ER-alpha [molecules/25 ng RNA]1e+3 1e+4 1e+5 1e+6

ER

-bet

a [m

olec

ules

/25

ng R

NA

]

1e+2

1e+3

1e+4

1e+5

Col 22 vs Col 26: 977883,75

Col 22 vs Col 26: 201818,75

Col 22 vs Col 26: 5421,71

Col 22 vs Col 26: 79348,75

Col 22 vs Col 26: 100003,75

Col 22 vs Col 26: 60479,88

Col 22 vs Col 26: 145339,88

Col 22 vs Col 26: 6035,36

Col 22 vs Col 26: 12027,5

Col 22 vs Col 26: 204646,5

Col 22 vs Col 26: 4061,57

Col 22 vs Col 26: 4591,62

Col 22 vs Col 26: 1535,57

Col 22 vs Col 26: 10212,38

Col 22 vs Col 26: 4160,5

ratio = 1

uterusliverlunglongissimus dorsihind leg musclesshoulder musclesneck musclesheartspleenmammary glandrumenabomasumjejunumkidney medullakidney cortex

Estrogen receptors (ERα & ERβ) expression pattern in cattle tissues

Pfaffl et al., APMIS 2001

Relative QuantificationThe mRNA expression is relative to WHAT ???

• relative to a non treated control

• relative to a time point zero

• relative to another gene of interest

• relative to the mean expression of a target gene

• relative to an universal calibration curve

• relative to the expression of one constant expressed HKGGAPDH, tubulins, various actins, albumins, cyclophilin, micro-globulins, histone subunits, ribosomal units (18S or 28S rRNA), ………….etc.

• relative to a HKG Index containing more HKGs (> 3)geNorm (Vandesompele et al., Genome Biology, 2002)BestKeeper (Pfaffl et al.; Biotechnology Letters 2004)Normfinder (Andersen et al., Cancer Research 2004)statistical modeling (Szabo et al., Genome Biology 2004)

• etc. ???

Normalisation strategiesAccording to known amounts of extracted RNA

(RIN quality check, molecules/ng RNA; ag transcript/ng RNA)

According to mass or volume of extracted tissue(molecules/mg tissue; ag transcript/mg tissue; transcript/cells)

According to one known and NOT regulated HKGGAPDH, tubulins, actins, albumins, cyclophilin, micro-globulins, histone subunits, ribosomal units (18S or 28S rRNA),.............

According to a HKG Index containing more HKGs (> 3)geNorm (Vandesompele et al. 2002, Genome Biology)Accurate normalization of real-time quantitative RT-PCR data by geometric veraging of multiple internal control genes.

BestKeeper (Pfaffl et al. 2004; Biotechnology Letters 2004)Determination of most stable housekeeping genes, differentially regulated target genes and sample integrity: BestKeeper Excel based tool using pair-wise correlations.

Normfinder (Andersen et al., 2004 Genome Biology)Normalization of Real-Time Quantitative Reverse Transcription-PCR Data: A Model-Based Variance Estimation Approach to Identify Genes Suited for Normalization.

Statistical modeling for selecting housekeeper genes(Szabo et al. 2004, Genome Biology)

Relative Quantification in real time qRT-PCR

without real-time PCR efficiency correction

2 (-∆∆ CP)

relative quantification

external calibration

curve withoutany referenece

gene

normalisationvia one

reference gene

via reference gene index

>3 HKG

ROX

“Delta-delta method” for comparing relative expression results between treatments in real-time PCR

presented by PE Applied Biosystems (Perkin Elmer, Forster City, CA, USA)

ABI Prism 7700 Sequence detection System User Bulletin #2 (2001)

Relative quantification of gene expression.http://docs.appliedbiosystems.com/pebiodocs/04303859.pdf

expression ratio = 2

- [ ∆CP sample - ∆CP control]

- ∆∆CPexpression ratio = 2

Livak KJ, Schmittgen TD. (2001) Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2 (- delta deltaC(T)) method.Methods. 2001 25(4): 402-408.

“Delta-delta method” for comparing relative expression results between treatments in real-time PCR

presented by ABI (Applied Biosystems Inc.)

assumptions: E = 2 & ∆CP is constant over a wide range

expression relative to the time point zero & normalised by a HKG:TNFα mRNA expression in cultured leukocytes after LPS stimulation:

white blood cells [WBC] vs. somatic milk cells [SMC] isolated blood monocytes vs. isolated milk macrophages

PCR Efficiency = 2 n = 6 mean ± sem

TNFalpha

Time[h]

0 2 4 6 8

TNFa

lpha

[del

taC

P]

0

2

4

6

8 WBCSomatic Cells

***

***

******

***

+++

++

+

+

******

***

***

*

+++

+++

TNFalpha

Time[h]

0 2 4 6 8

TNFa

lpha

[del

taC

P]

0

2

4

6

8 MonocytesMacrophages

*** +***

++ ******

+++***

***+++

***

+++ ***

+++

Prgomet et al., 2004

Relative Quantification in real time qRT-PCR

without real-time PCR efficiency correction

2 (-∆∆ CP) REST, qGeneLC software

with real-time PCR efficiency correction

relative quantification

external calibration

curve withoutany referenece

gene

normalisationvia one

reference gene

via reference gene index

>3 HKG

ROX

Tissue “background” interfere with real-time PCR efficiency and amplification fidelity

IGF-1 mRNA amplification in three cattle tissues

liver

m. splenius m. gastrocnemius

Efficiency variation in real-time RT-PCR

Roche Diagnostics, LC rel. Quantification software, March 2001

PCR inhibitors:Hemoglobin, Urea, Heparin

Organic or phenolic compoundsGlycogen, Fats, Ca2+

Tissue matrix effectsLaboratory items, powder, etc.

PCR enhancers:DMSO, Glycerol, BSA

Formamide, PEG, TMANO, TMAC etc.Special commercial enhancers:

Gene 32 protein, Perfect Match, Taq Extender, AccuPrime, E. Coli ss DNA binding

real-time PCRefficiency

NA degradation = RIN

PCR reaction components

DNA concentration

tissue degradation

unspecificPCR products

hardware:PCR platform & cups

lab management DNA dyes cycle conditions

Relative quantification of a target gene versus areference gene (housekeeping gene)

Single data (n = 1) e.g. array results:

Etarget∆CPtarget (control - sample)

relativeexpression =

Ereference∆CPref (control - sample)

Pfaffl, Nucleic Acids Research 2001

relativeexpression

( Eref )CPSample

( Etarget )CPSample

÷( Eref )

Calibrator

( Etarget )CPCalibrator

=

Roche Diagnostics, LC relative Quantification software, March 2001

CP

Relative quantification of a target gene versus areference gene (housekeeping gene)

single data (n = 1) e.g. array results:

Etarget∆CPtarget (control - sample)

relativeexpression =

Eref∆CPref (control - sample)

Pfaffl, Nucleic Acids Research 2001

multiple data (1 < n < 100) e.g. experimental groups via REST-XL©:

Etarget∆CPtarget (MEAN control – MEAN sample)

relativeexpression =

Eref∆CPref (MEAN control – MEAN sample)

Pfaffl et al., Nucleic Acids Research 2002

http://rest.gene-quantification.info

Influence of total RNA quality, quantity and purity on qRT-PCR resultstotal RNA extracted bovine WBC analyzed in Bioanalyzer 2100

ladder RIN: 9.5

RIN: 5.6 RIN: 2.8

V. Walf, S. Huch, MW. Pfaffl, 2005

RIN in different bovine tissues (Box Plot)R

IN

0

2

4

6

8

10

liver (n = 22)heart (n = 17)spleen (n = 17)lung (n = 22)rumen (n = 23)reticulum (n = 26)omasum (n = 17)abomasum (n = 17)ileum (n = 17)jejunum (n = 20)colon (n = 19)caecum (n = 16)mes. lymphnode (n = 26)

S. Fleige & MW. Pfaffl, 2005

RIN liver 1st and 2nd extraction

0

2

4

6

8

10

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

sample number

RIN

Leber 1.Extraktion Leber 2.Extrakion

Enzymatic or UV degradation of tissue extracted total RNA

RIN 10.0 9.2 9.2 8.6 8.1 7.8 7.2 6.7 6.0 5.0 4.4 4.0

RIN 9.5 RIN 2.8

bovine WBC total RNA

∆ CP

Influence of total RNA quality on qRT-PCR results

y = -0,206x + 7,6396R2 = 0,5363

y = -0,887x + 21,398R2 = 0,7165

y = -0,7211x + 27,077R2 = 0,7682

0,0

5,0

10,0

15,0

20,0

25,0

30,0

0 1 2 3 4 5 6 7 8 9 10

RNA Integrity Number

CP

= T

ake

Off

Po

int

28S

beta actin

IL-1

bovine ileum total RNA

28S18S

β-ActinIL-1β

bovine ileum: CP = TOP

y = -0,2036x + 8,6557R2 = 0,7612

y = -2E-15x + 6,00R2 = 1,0

y = -0,1461x + 13,565R2 = 0,4903

R2 = 0,8119

0,0

5,0

10,0

15,0

20,0

25,0

0 1 2 3 4 5 6 7 8 9 10

RNA Integrity Number

CP

= Ta

ke O

ff Po

int

18S 28S beta-actin IL-1 Linear (18S) Linear (28S) Linear (beta-actin) Linear (IL-1)

18S

28S

beta actin

IL-1

CP = constant in relative quantification

y = -0,2583x + 21,787∆

28S n.s.Interleukin-1beta P < 0.001

beta-Actin P < 0.0118S P < 0.001

Influence of total RNA quality on qRT-PCR CP (Ct)IL-1: Crossing Point

14,0

16,0

18,0

20,0

22,0

24,0

26,0

28,0

30,0

32,0

34,0

0 1 2 3 4 5 6 7 8 9 10RNA Integrity Number

Cro

ssin

g P

oint

Reticulum (E) Lymph nodes (E) Lymph nodes (P) Colon (P) Lung (E)Corpus luteum (P) Caecum (P) Spleen (P) Abomasum (P)

Influence of total RNA quality on qRT-PCR efficiency

28S: Amplification

1,5

1,6

1,7

1,8

1,9

2,0

0 1 2 3 4 5 6 7 8 9 10RNA Integrity Number

PCR

Eff

icie

ncy

Reticulum (E) Lymph nodes (E) Lymph nodes (P) Colon (P) Lung (E)Corpus luteum (P) Caecum (P) Spleen (P) Abomasum (P)

Advantages:• each sample derived from identical total RNA (one-step quantification) or cDNA

source (two-step quantification) = identical inhibitors and enhancers• tissue specific matrix effects are given, but not existent in relative quantification• total RNA quantity, RT efficiency and cDNA quantity are minor relevant • effect of total RNA quality and influence on qRT-PCR have to be tested in detail• optimized assays are highly reproducibility (variations only given by researcher, kit,

robot and real-time platform)• each sample will be normalised with HKG or better by a HKG INDEX• no production of a calibration curve: e.g. cloning, linearization or purification of

calibration curve material• no data conversion: molecules, concentrations or molarities• => direct procedure: measured data were inserted in mathematical model• => relative quantification by software applications, e.g. qGene, REST, LightCycler• => automatic calculation => bio-informatics session

Disadvantages:• No information about the effective molecule concentration present in the tissues

(molecules/mg tissue; molecules/ng RNA; molecules /cell, etc.)

Relative quantification using efficiency corrected calculation

Thank you team !Thank you for your attention !